. Transient ischemic attacks (TIA) are common and important warning signs of stroke. Fifteen percent of stroke patients first present with a TIA, so preventing strokes after TIA could significantly reduce the total stroke burden. Little evidence has been available about the natural history, risk factors for stroke, and appropriate management after TIA, leading to tremendous practice variability. In a prior study, the applicant studied short-term prognosis in a cohort of TIA patients diagnosed in emergency departments of a large managed care organization in Northern California. Eleven percent of 1707 patients in the cohort had a stroke within 3 months of presentation, and nearly two-thirds of these occurred within 1 week of the index TIA. Age, diabetes mellitus, duration of the event, and signs or symptoms of weakness with the spell were strong, independent risk factors for stroke. The high risk of stroke after TIA underlines the potential benefit of urgent interventions, and the identification of risk factors suggests these interventions could be targeted efficiently. This proposal extends this study to provide tools and evidence for optimizing the management of transient ischemic attacks while extending the applicant's expertise in health services research in cerebrovascular disease. The proposed research program includes the following studies: 1) Development and validation of simple clinical prediction rules to identify patients at minimal and high short-term risk of stroke after TIA; 2) Evaluation of management practices in TIA using a new multivariable method, termed two-level analysis; 3) Evaluation of imaging study results as prognostic indicators in TIA patients. This research will provide important results for clinicians and policy makers, and will foster the applicant's development as a health services researcher in cerebrovascular disease.